July 2008 Update on Global Temperature – Additional HadCrut Analysis

Well, I finally got to my spreadsheet and analysis on the HadCrut data! It’s an interesting review for a few reasons: (1) as another temperature measure based on surface-station data rather than satellite data, it provides a check against GISS and NOAA, (2) it goes back another 30 years (who knows how much that can be trusted, but it is what it is), and (3) HadCrut doesn’t publish its methodology, so it’s a total black box. NOAA and NASA may be as confusing as all get-out to try and figure out, but at least it is publicly available, and you do have guys like Steve McIntyre at ClimateAudit actually figuring out how to replicate the NASA anomalies. HadCrut takes the position that it’s their method and wants to protect it. Not exactly high on the scale of credibility assessment, but it doesn’t mean it’s not a good estimate, either. It’s just one of those things you need to know as you review the information.

I already gave a quick review of the HadCrut data earlier in the month, so I won’t go into great detail. I thought it may be interesting to throw up a couple charts that demonstrate the state of the HadCrut trends, though, so I put them out here for your amusement.

Overall slope since 1850 on the HadCrut data shows average warming at less than half a degree Celsius per Century.  (0.436 degrees, to be exact)

The current period of cooling on HadCrut extends back to April 1997, which is actually further back than the Satellite measures, and nearly four more years than the NOAA and GISS surface measures.

The last 60 months shows a steep slope downward.

The slope has continually become more negative since early 2004.

We can see in the cyclical variations of the 60-month trend over time that since around 1950, the low points of the trend reached higher minimum slope points, but the most recent measure has broken through that trend. When that same thing happened around 1945, the 60-month slope continued to plunge.

Predicted Anomalies For July
As with my GISS and NOAA models, I have built in a slope-weighting mechanism to try and predict the changes in slope trends and the associated anomalies that would be required. Based on this model, I am predicting a July anomaly in the HadCrut data of between 31.5 and 33.4. The wider range of predicted values across the different methods is 20.6 to 36.2. The 20.6 is somewhat of an outlier.